*Replication codes for                        
*Extended Dependence: Trade, Alliances, and Peace
*Forthcoming, The Journal of Politics            
                                                
*Frederick R. Chen                               
*University of Wisconsin-Madison                 
*frederick.chen@wisc.edu                         
                                                
*August 9, 2019                                  


use "D:\Extended Dependence\2019JOPreplication\replication.dta", clear

sort ddyad year
xi i.year

*Main analysis

*Table 1 Extended dependence and international conflict initiation

*Model 1
relogit midsidea dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model 2
relogit midsidea extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model 3
relogit midsidea extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model 4
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*No Fatal MID in 1968

*Model 5
relogit midsideafatal dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3  _Iyear_1952 - _Iyear_1967 _Iyear_1969 -_Iyear_2010 if year!=1968, cluster(ddyad)

*Model 6
relogit midsideafatal extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model 7
relogit midsideafatal extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model 8
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Table 2 Relative risks of MIDs

*All MIDs
logit midsidea extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3, cluster(ddyad)
margins, at(extdeporig=(0, 6.82, 26.18, 45.54)) atmeans post

*Fatal MIDs
logit midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3, cluster(ddyad)
margins, at(extdeporig=(0, 6.82, 26.18, 45.54)) atmeans post

*Table 3 Nonaggression-pact dependence and international conflict initiation

*Model 9
relogit midsidea napdep dyaddep0 napcapshare comnapno dyadnap c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model 10
relogit midsideafatal napdep dyaddep0 napcapshare comnapno dyadnap c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Online appendix

*Appendix A Summary statistics

sum extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist

corr midsidea midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist

*Appendix B Potential bias from missing bilateral trade data

relogit midsidea extdepie dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

relogit midsidea extdepie dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

relogit midsidea extdepie dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

relogit midsideafatal extdepie dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

relogit midsideafatal extdepie dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

relogit midsideafatal extdepie dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Appendix C Dyadic trade interdependence

use "D:\Extended Dependence\2019JOPreplication\nondirected.dta", clear

*Model C1
relogit mid depl polityl polityh atopally capshare majpow cont lndist pcyrmid1 pcyrmid2 pcyrmid3, cluster(dyad)

*Model C2
relogit mid depl0 polityl polityh atopally capshare majpow cont lndist pcyrmid1 pcyrmid2 pcyrmid3, cluster(dyad)

*Model C3
relogit midfat depl polityl polityh atopally capshare majpow cont lndist pcyrmidfat1 pcyrmidfat2 pcyrmidfat3, cluster(dyad)

*Model C4
relogit midfat depl0 polityl polityh atopally capshare majpow cont lndist pcyrmidfat1 pcyrmidfat2 pcyrmidfat3, cluster(dyad)

*Appendix D Robustness checks

use "D:\Extended Dependence\2019JOPreplication\replication.dta", clear

sort ddyad year
xi i.year

*Table D1 Using politically relevant dyads only

*Model D1.1
relogit midsidea dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if polrel==1, cluster(ddyad)

*Model D1.2
relogit midsidea extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if polrel==1, cluster(ddyad)

*Model D1.3
relogit midsidea extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if polrel==1, cluster(ddyad)

*Model D1.4
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if polrel==1, cluster(ddyad)

*Model D1.5
relogit midsideafatal dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3  _Iyear_1952 - _Iyear_1967 _Iyear_1969 -_Iyear_2010 if year!=1968 & polrel==1, cluster(ddyad)

*Model D1.6
relogit midsideafatal extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968 & polrel==1, cluster(ddyad)

*Model D1.7
relogit midsideafatal extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968 & polrel==1, cluster(ddyad)

*Model D1.8
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968 & polrel==1, cluster(ddyad)

*Table D2 Treating revisionist states as initiators

*Model D2.1
relogit midrev dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidrev1 pcyrmidrev2 pcyrmidrev3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D2.2
relogit midrev extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidrev1 pcyrmidrev2 pcyrmidrev3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D2.3
relogit midrev extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidrev1 pcyrmidrev2 pcyrmidrev3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D2.4
relogit midrev extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidrev1 pcyrmidrev2 pcyrmidrev3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*No Fatal MID in 1968 or 1988

*Model D2.5
relogit midrevfatal dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidrevfatal1 pcyrmidrevfatal2 pcyrmidrevfatal3  _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_1987 _Iyear_1989 - _Iyear_2010 if year!=1968 & year!=1988, cluster(ddyad)

*Model D2.6
relogit midrevfatal extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidrevfatal1 pcyrmidrevfatal2 pcyrmidrevfatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_1987 _Iyear_1989 - _Iyear_2010 if year!=1968 & year!=1988, cluster(ddyad)

*Model D2.7
relogit midrevfatal extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidrevfatal1 pcyrmidrevfatal2 pcyrmidrevfatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_1987 _Iyear_1989 - _Iyear_2010 if year!=1968 & year!=1988, cluster(ddyad)

*Model D2.8
relogit midrevfatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidrevfatal1 pcyrmidrevfatal2 pcyrmidrevfatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_1987 _Iyear_1989 - _Iyear_2010 if year!=1968 & year!=1988, cluster(ddyad)

*Table D3 Taking the sum of extended and dyadic dependence

gen depsum = extdeporig + dyaddep0

*Model D3.1
relogit midsidea depsum dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D3.2
relogit midsidea depsum dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D3.3
relogit midsideafatal depsum dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model D3.4
relogit midsideafatal depsum dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Table D4 Controlling for US allies and NATO members

*Model D4.1
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno c2usdp dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D4.2
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno c2nato dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D4.3
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno c2usdp dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model D4.4
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno c2nato dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Table D5 Including additional control variables

*Model D5.1
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno c2dpno sscore coigo dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D5.2
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno c2dpno sscore coigo dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Table D6 Removing potential outliers

*Model D6.1
relogit midsidea extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if extdeporig<=200, cluster(ddyad)

*Model D6.2
relogit midsidea extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if extdeporig<=200, cluster(ddyad)

*Model D6.3
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010 if extdeporig<=200, cluster(ddyad)

*Model D6.4
relogit midsideafatal extdeporig dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968 & extdeporig<=200, cluster(ddyad)

*Model D6.5
relogit midsideafatal extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968 & extdeporig<=200, cluster(ddyad)

*Model D6.6
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968 & extdeporig<=200, cluster(ddyad)

*Table D7 Using listwise deletion for dyadic dependence

*Model D7.1
relogit midsidea dyaddep dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D7.2
relogit midsidea extdeporig dyaddep dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D7.3
relogit midsidea extdeporig dyaddep dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D7.4
relogit midsideafatal dyaddep dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3  _Iyear_1952 _Iyear_1953 _Iyear_1955 - _Iyear_1967 _Iyear_1969 -_Iyear_2010 if year!=1968, cluster(ddyad)

*Model D7.5
relogit midsideafatal extdeporig dyaddep dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3  _Iyear_1952 _Iyear_1953 _Iyear_1955 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model D7.6
relogit midsideafatal extdeporig dyaddep dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3  _Iyear_1952 _Iyear_1953 _Iyear_1955 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Table D8 Using democratic dyad

*Model D8.1
relogit midsidea dyaddep0 dyaddp c2majpow demdyad cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D8.2
relogit midsidea extdeporig dyaddp c2majpow demdyad cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D8.3
relogit midsidea extdeporig dyaddep0 dyaddp c2majpow demdyad cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D8.4
relogit midsidea extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow demdyad cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 _Iyear_1952 - _Iyear_2010, cluster(ddyad)

*Model D8.5
relogit midsideafatal dyaddep0 dyaddp c2majpow demdyad cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3  _Iyear_1952 - _Iyear_1967 _Iyear_1969 -_Iyear_2010 if year!=1968, cluster(ddyad)

*Model D8.6
relogit midsideafatal extdeporig dyaddp c2majpow demdyad cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model D8.7
relogit midsideafatal extdeporig dyaddep0 dyaddp c2majpow demdyad cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Model D8.8
relogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow demdyad cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 _Iyear_1952 - _Iyear_1967 _Iyear_1969 - _Iyear_2010 if year!=1968, cluster(ddyad)

*Table D9 Including directed-dyad fixed effects

xtset ddyad year

*Model D9.1
xtlogit midsidea extdeporig pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 i.year, fe vce(oim)

*Model D9.2
xtlogit midsidea extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 i.year, fe vce(oim)

*Model D9.3
xtlogit midsidea extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsidea1 pcyrmidsidea2 pcyrmidsidea3 i.year, fe vce(oim)

*Model D9.4
xtlogit midsideafatal extdeporig pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 i.year, fe vce(oim)

*Model D9.5
xtlogit midsideafatal extdeporig dyaddep0 dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 i.year, fe vce(oim)

*Model D9.6
xtlogit midsideafatal extdeporig dyaddep0 dpcapshare comdpno dyaddp c2majpow c1dem c2dem cont dist pcyrmidsideafatal1 pcyrmidsideafatal2 pcyrmidsideafatal3 i.year, fe vce(oim)

*Appendix E Additional evidence for the causal mechanisms

*Table E1 Trade between the challenger and joiner on the target side

clear
use "D:\Extended Dependence\2019JOPreplication\mech1.dta"
xtset dyad year

*Model E1.1
xtreg trade mid tradelag i.year, fe 

*Model E1.2
xtregar trade mid, fe

*Model E1.3
xtreg trade midfat tradelag i.year, fe 

*Model E1.4
xtregar trade midfat, fe

*Table E2 Trade between the challenger and allies of the target

clear
use "D:\Extended Dependence\2019JOPreplication\mech3.dta"
xtset dyad year

*Model E1.1
xtreg trade mid tradelag i.year, fe 

*Model E1.2
xtregar trade mid, fe

*Model E1.3
xtreg trade midfat tradelag i.year, fe 

*Model E1.4
xtregar trade midfat, fe